Leica M Monochrom DNG raw question

[250swb]

Advertisement (gone after registration)

If yes, the converter has to cut off some of the information in order to export a TIFF/JPG.

 

I'd be interested to know why the converter has to cut of some information with a TIFF file, its not like a TIFF is limited to a size or anything, and you have a histogram with the relevent numbers to look at in the RAW converter, and you have a histogram to look at with the relevent numbers in post processing. You would see if anything went missing.

 

Steve

[jpk]

I don't know if it's true, but: if DNG can store more (or finer or better or more accurate) information than a TIFF and if the image in question uses this more-of-information - only in this case the converter would have to cut off... Right?

[koferk]

interesting article-

Color vs. Monochrome Sensors

"Unlike with color, monochrome sensors also do not require demosaicing to create the final image—

the values recorded at each photosite effectively just become the values at each pixel." - exactly what i meant..

 

One should also consider whether the quality of a monochrome sensor outweighs the flexibility of a color sensor. - no "raw" flexibility in monochrome, its the same as not having raw but only having a tif..

 

the first quote is the more suited, its just says exactly what i mean.

[jaapv]

I don't know if it's true, but: if DNG can store more (or finer or better or more accurate) information than a TIFF and if the image in question uses this more-of-information - only in this case the converter would have to cut off... Right?

 

The operative word being "if" and "relevant" should be added as well. It is obvious the exported TIFF will not contain the embedded JPG or DNG header, for instance. But as a number of raw files are based on TIFF/EP, including DNG, the point is moot anyway.

We must differentiate between raw data coming from a Bayer filtered sensor and files coming from the `Monochrom. The extra steps needed in conversion to create the colour by interpolation and demosaicing are raw-developer specific algorithms and can lead to different results. The conversion from the Monocrom-specific LINEAR DNG is different and not supported by all raw developers.

 

.Note that raw conversion consists of eight to fourteen steps, only three of which have to do with colour.

 

"Linear DNG": A rarer variety of DNG, containing RGB image data (or not just RGB!) arranged in a rectilinear format. (Most practical examples of Linear DNG hold RGB image data, but it isn't a constraint, and 4 or more colours are allowed.[.....]This RGB image data may have come from demosaiced raw image data, or from another source such as TIFF or JPEG or something else. (There is little difference between "demosaiced" and "never mosaiced") *

.

 

 

 

*quote from Barry Pearson's website

[wattsy]

So the remainig question is: can the TIFF-like saved RAW-data inside the DNG contain more information than the exported TIFF? If yes, the converter has to cut off some of the information in order to export a TIFF/JPG

 

I don't think the DNG can contain more information (and I certainly don't think there is any "vector" like information in it) but it may contain much flatter (more linear) information. It is the introduction of a curve, black and white points and the making of contrast (and sharpening) decisions that may involve cutting off "some of the original information in order to export a TIFF/JPG".

[wattsy]

no "raw" flexibility in monochrome, its the same as not having raw but only having a tif.

 

Yes, I think this is probably true if the comparison is with a linear unsharpened TIFF file (although Sandy makes an interesting point about the DNGs containing blacks that are "blacker than black":)).

[koferk]

Advertisement (gone after registration)

"eight to fourteen steps" - where do you guys bring this stuff from?

is it 8 or 14 ? or maybe 12 ? but more important - what are they???

 

there is no such thing as blacker then black, the same way there is no such thing as "more" yellow then yellow or more cheese then cheese.

 

all there is to it is a photosite, and a pixel, which have the same data value. one represent that data in raw form and the other in bitmap form (tif for example). the way i see it, theres nothing more to it. its the same data packed in a different form. that's all.

 

(i guess i still don't get it)

[jaapv]

There are eight essential steps and six optional ones. Start understanding stuff before you pontificate.

Your black remark is silly. Black point is defined in raw by the camera firmware but can be shifted in raw conversion and postprocessing.

And yes, there can be more yellow than yellow, specifically in LAB where one can define colours that do not exist - but that is wholly beyond the scope of this discussion.

[koferk]

Im sorry for my style.. my english isn't the best..

could you please say what are those 14 steps ?

thanks.

 

I agree that "Black point is defined in raw by the camera firmware but can be shifted in raw conversion and postprocessing."

that still doesn't make no black blacker then any other black.. they are all blacks..

and also, the fact that something can be done in raw, doesn't mean its raw only.. a simple example can be crop. do it in raw, do it in tif, it wont matter.

 

i have the feeling that all i read here is explanations regarding regular raw and no explanations for the "special" (simple!) monochrom raw.

[jaapv]

I give up...

Maybe it is best for you to accept that reality is that digital cameras produce either a proprietary raw file like NEF,CRW,DNG, etc which must be converted to a format that your computer can handle or an universal, but more limited, JPEG file. Or even both.

 

Even dedicated companies like DxO (who really know about raw, conversions and file formats, I can assure you) will only work with camera output raw, manipulated by firmware and all, not with sensor data for their measurements.

[koferk]

that is simply not true.

many of nikon's dslr's can shoot tif format, in camera.

but it also unrelated.

 

can someone tell me what are the 14 steps the raw converter does ?

thanks.

[jaapv]

I thought you were such a Wikipedia fan?

 

 

To be viewed or printed, the output from a camera's image sensor has to be processed, that is, converted to a photographic rendering of the scene, and then stored in a standard raster graphics format such as JPEG. This processing, whether done in-camera or later in a raw-file converter, involves a number of operations, typically including:

 

decoding – image data of raw files are typically encoded for compression purpose, but also often for obfuscation purpose (e.g. raw files from Canon or Nikon cameras)[citation needed].

defective pixel removal – replacing data in known bad locations with interpolations from nearby locations

white balancing – accounting for color temperature of the light that was used to take the photograph

demosaicing – interpolating the partial raw data received from the color-filtered image sensor into a matrix of colored pixels.

noise reduction – trading off detail for smoothness by removing small fluctuations

color translation – converting from the camera native color space defined by the spectral sensitivities of the image sensor to an output color space (typically sRGB for JPEG)

tone reproduction – the scene luminance captured by the camera sensors and stored in the raw file (with a dynamic range of typically 10 or more bits) needs to be rendered for pleasing effect and correct viewing on low-dynamic-range monitors or prints; the tone-reproduction rendering often includes separate tone mapping and gamma compression steps.

compression – for example JPEG compression

Note that demosaicing is only performed for CFA sensors; it is not required for 3CCD or Foveon X3 sensors.{..and non-Bayer-filtered sensors}

 

Cameras and image processing software may also perform additional processing to improve image quality, for example:

 

removal of systematic noise – bias frame subtraction and flat-field correction

dark frame subtraction

optical correction – lens distortion correction, vignetting correction, and color fringing correction

contrast enhancement

increasing visual acuity by unsharp masking

dynamic range compression – lighten shadow regions without blowing out highlight regions

 

When a camera saves a raw file it defers most of this processing; typically the only processing performed is the removal of defective pixels (the DNG specification requires that defective pixels are removed before creating the file). Some camera manufacturers do additional processing before saving raw files; for example, Nikon has been criticized by astrophotographers for applying noise reduction before saving the raw file.

 

 

 

 

And:

http://en.wikipedia.org/wiki/Color_image_pipeline

[koferk]

very interesting!

thanks for keep on going here

 

so i read the quote, lets go over those eight essential steps for a moment for a check up -

1. decoding - for compression which is just a matter of file size. (nikon and canon seems to do some kind obfuscation but that's surly not essential if only some camera makers do it)

2. defective pixel removal - my favorite and the best thing i've learned so far, but then again, assuming your camera is fine, it wouldn't matter.

3. white balancing - as someone here mentioned and i've learned, there's no white balance for monochrome.

4. demosaicing - on this we all agree, no filter for us

 

i will continue till number 8, even though everything after demosaicing is either demosaicing related, or color related.

 

5. noise reduction - mosaic filter related. (please correct me im not sure how this actually works, i can only imagine its related)

6. color translation - no color for us.

7. tone reproduction - a color thing ? filter stuff ? not sure.

8. compression - jpeg stuff.

 

after number 8 they write it clear: "Note that demosaicing is only performed for CFA sensors; it is not required for 3CCD or Foveon X3 sensors.{..and non-Bayer-filtered sensors}"

so it seems after the demosaicing its all related to it in this way or another.

 

to sum it up - im still high above

please feel free to keep correct me, i've learned a lot so far, thanks!!

[jaapv]

One thing: if a camera would use a sensor free of defects, it would cost in excess of 25.000 $

[koferk]

don't you think 8g should be enough ?

[jaapv]

Yes - the point being that all cameras use sensors with a number of dead/hot pixels that must be mapped out in firmware. The lower percentage the higher the price. Fault-free sensors are incredibly expensive and never used for "normal" photographic applications. So defective pixel removal does not "not matter", it is an essential step, no matter how "fine" your camera is.

 

As for initial compression, it s not just a matter of file size in the sense of filling the buffer, it reduces processing time and speeds up the camera. For instance Leica initially did not compress the M8 files and found that the prototype camera slowed down unacceptably, so they had to introduce lossy compression.

 

And then you left out noise reduction, which all cameras do for noise that cannot be eliminated or reduced in post processing at least, and often far more, like Nikon.

 

Tone reproduction, an obvious essential and even more so for the Monochrom.

 

One other thing you are missing out on here is specific for the M series, which is correction of the corners for wide-angle lenses.

 

These are operations that are impossible/very hard or time-consuming to do in post processing, so are far better handled in raw conversion and firmware.

 

Btw, you claim that Nikons can output TIFF files. I looked up the D600, as I am not familiar with Nikon, and that one doesn't. Haven't you got it confused with NEF, which is TIFF-based?

once

So again, you will have to live with either DNG or JPG and no amount of arguing will get you anywhere. And once again,there is no technical reason not to use raw in DNG form and many advantages, so the whole argument is a quite silly.

[koferk]

its important to mention this is only a logical question for me, im just trying to understand and not to "do" anything, at least not at the moment.

 

secondly, the Nikon D800 for example can shoot TIFF

Nikon D800 D-SLR Camera | High Dynamic Range Camera

 

now for our main subject, you wrote "These are operations that are impossible/very hard or time-consuming to do in post processing, so are far better handled in raw conversion and firmware."

 

and so my question is, which specific one is impossible, and more importantly, why is it impossible in your opinion ? meaning, as to get the same results, time and computing power is irrelevant for my question.

[jaapv]

Sorry, my friend. End of the line. Time to start doing your own research....

[koferk]

bummer.. well at least i tried, as hard as i could

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[250swb]

The effort made to find the shift key and use capital letters is symptomatic of how hard you tried, so if you found that difficult I don't doubt everything else was as well.

 

Steve